Raw material conveying device for food processing

Abstract

The invention discloses a raw material conveying device for food processing. The raw material conveying device comprises a conveying body, a feeding device, a discharging storage box, an industrial vision intelligent detection assembly and a controller. The conveying body is composed of a recognition section, a flow dividing device, a detection section, a confluence section and other sections of conveying belts. Raw materials are graded according to the particle size through the flow dividing device, the types, impurities and defects of the raw materials are automatically recognized through the visual detection assembly, the detection conditions are optimized through the thinning mechanism, and unqualified products are removed through the automatic removing device. The controller is integrated with an image recognition and linkage control module, and real-time monitoring, data recording and intelligent scheduling of the conveying process are achieved. The device effectively solves the problems that traditional conveying depends on manpower, detection coverage is low, response lags behind and tracing is difficult, and the automation and intelligence level of food raw material conveying and the product quality safety are remarkably improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of food processing equipment, and particularly relates to a raw material conveying device for food processing. BACKGROUND

[0002] In the field of food processing, raw material conveying is a key link connecting raw material storage and production processing, which directly affects product quality and production efficiency. At present, the raw material conveying devices widely used by food processing enterprises, such as belt conveyors, screw conveyors and pipeline conveyors, still have many technical deficiencies in actual application, and it is difficult to meet the needs of modern food production for safety, accuracy and intelligence.

[0003] Firstly, the existing conveying device mainly relies on manual sampling inspection for raw material quality detection, which has obvious limitations: on the one hand, manual detection is low in efficiency and high in cost, and it is difficult to realize full-process and full-coverage detection. For small impurities such as metal scraps, plastic impurities and stones in granular or powdery raw materials, as well as abnormal states such as damp clumping, discoloration and deterioration of raw materials, it is easy to miss detection; on the other hand, manual judgment is affected by subjective experience, fatigue degree and other factors, and the detection standard is not unified, which may lead to unqualified raw materials entering the subsequent processing link and causing food safety hazards.

[0004] Secondly, the traditional conveying device lacks effective real-time monitoring means, and the response to abnormal conditions in the conveying process is lagging: for example, raw materials may be blocked, off-loaded or leaked during conveying, which needs to be discovered by manual and then processed, often leading to raw material waste, equipment damage, and even affecting the continuous operation of the production line; at the same time, the control of conveying flow mainly depends on experience adjustment, and it is difficult to dynamically adjust according to the actual conveying amount of raw materials, which easily leads to excessive or insufficient supply, affecting the stability of the subsequent processing technology.

[0005] In addition, the existing conveying device has not established a perfect data recording and tracing system, and the key information such as raw material batch information, quality detection data and equipment running state in the conveying process cannot be effectively retained. When product quality problems occur, it is difficult to quickly trace the source, conveying path and related process parameters of the problem raw materials, which is not convenient for the investigation and responsibility definition of quality problems, and also difficult to meet the requirements of food production safety regulations on traceability management.

[0006] Therefore, it is a key requirement to develop an intelligent food raw material conveying device capable of realizing accurate detection of raw material quality, real-time monitoring of conveying process and data tracing throughout the process. SUMMARY

[0007] In order to overcome the problems of scheduling not timely, response lag, unreasonable route planning in the prior art power grid data management, the application discloses a raw material conveying device for food processing, which can realize efficient operation of mortar tank trucks, improve mortar transportation efficiency and reduce resource waste.

[0008] To solve the above technical problems, the technical scheme adopted by the application is as follows: a raw material conveying device for food processing, comprising a conveying main body, a feeding device, a discharging storage box, an industrial visual intelligent detection assembly and a controller; the feeding device and the discharging storage box are respectively arranged at two ends of the conveying main body, the conveying main body is composed of a plurality of conveying belts, including an identification section conveying belt, a shunt conveying device, a detection section conveying belt and a confluence section conveying belt, the feeding device comprises a feeding hopper and a screw conveyor, the feeding end of the screw conveyor extends into the bottom of the feeding hopper, and the discharging end of the screw conveyor is located above the identification section conveying belt; the surface of the detection section conveying belt is provided with baffles at fixed intervals along the conveying direction, fixed baffle plates are arranged on the two sides of each detection section conveying belt to block the two sides of the baffles, and a raw material compartment is formed between the two baffles; The shunt conveying device comprises a shell frame, a sliding plate and sieve plates, the sliding plate and the sieve plates are arranged obliquely, the sliding plate is fixed at the bottom, a plurality of sieve plates are arranged above the sliding plate, the sieve plates shunt food raw materials of different particle sizes, and the sliding plate and the sieve plates convey food raw materials into different detection section conveying belts through sliding channels respectively; The industrial visual intelligent detection assembly is arranged above the detection section conveying belt, and a thinning mechanism and an automatic rejection device are arranged at the industrial visual intelligent detection assembly; The thinning mechanism is used for thinning food raw materials, so that the height of the food raw materials does not exceed a threshold value and the visual detection is not affected; The industrial visual intelligent detection assembly comprises a first identification camera, an industrial camera and a mounting bracket, the first identification camera is arranged above the identification section conveying belt through the mounting bracket, the first identification camera is used for identifying the type and granularity of food raw materials, and the industrial camera is arranged on the detection section conveying belt through the mounting bracket, the industrial camera is used for shooting food raw materials on the detection section conveying belt and identifying impurities and defective raw materials; The driving motor of each conveying belt, the industrial visual intelligent detection assembly, the flattening mechanism and the automatic rejection device are electrically connected with the controller, the controller is provided with an image recognition module, a data processing module, a linkage control module, an interaction module and a data storage module, the image recognition module is used for detecting and identifying the images captured by the first identification camera and the industrial camera, the data processing module is used for processing and logically judging the data, the linkage control module receives signals and drives the execution mechanism, and is used for controlling the operation of each conveying belt and the shooting timing of each camera; a touch screen for interaction is arranged at the controller, the interaction module is used for controlling the display of the touch screen and receiving the input of the touch screen; the data storage module is used for storing food raw material data, a feature library, system pre-calibration parameters and defect tolerance rules, and the food raw material data includes the appearance of qualified food raw materials and the size of qualified food raw materials; The automatic rejection device is used for rejecting the impurities and defective raw materials identified by the industrial camera.

[0009] Further, the industrial visual intelligent detection assembly further comprises an annular RGBW programmable light source; the industrial camera transmits the collected images to the image recognition module, the annular RGBW programmable light source can switch four different color light compensation, and can also switch the light source point position to change the local illumination angle, and the linkage control module controls the operation of the annular RGBW programmable light source.

[0010] Further, the working process of the raw material conveying device for food processing includes: Step S1, turn on the system power supply, select the type of food raw material to be detected on the touch screen, and apply the corresponding food raw material data stored in the data storage module; Step S2, pour the food raw material into the feeding hopper, the spiral conveyor conveys the food raw material to above the identification section conveying belt, and the food raw material falls onto the identification section conveying belt; Step S3, the first identification camera captures images, identifies the type of food raw material, and judges whether it is consistent with the setting, if the type is different from the type set in step S1, an alarm is sent through the touch screen, the food raw material is conveyed into the shunt conveying device by the identification section conveying belt, the screen plate of the shunt conveying device screens the food raw material according to the particle size, and the food raw material is sent into the detection section conveying belt located at different layers respectively; Step S4, when the raw material compartment of the detection section conveying belt moves below the industrial visual intelligent detection assembly, the linkage control module controls the detection section conveying belt to stop running; Step S5, the flattening mechanism flattens the raw material in the raw material compartment that moves below the industrial visual intelligent detection assembly; Step S6, the industrial camera of the industrial visual intelligent detection assembly collects images and performs visual detection, the image recognition module detects whether there is impurity or unqualified food raw material, and feeds back the size and coordinates to the linkage control module; Step S7, the linkage control module controls the automatic rejection device to reject the impurities and defective raw materials identified by the industrial camera; Step S8, the linkage control module controls the detection section conveyor belt to move a fixed distance, so that the next raw material compartment moves under the industrial visual intelligent detection assembly, and the raw material compartment that has completed detection moves to the end of the detection section conveyor belt. The food raw materials on the detection section conveyor belt of each layer fall into the convergence section conveyor belt and are uniformly conveyed to the discharge storage box.

[0011] Further, the specific process of collecting images by the industrial camera of the industrial visual intelligent detection assembly includes: Step S6-1, the lamp beads of the annular RGBW programmable light source are evenly divided into eight blocks at equal angles, and the lamp beads of one block are lit in turn, and an image is collected each time the block is switched; Step S6-2, the annular RGBW programmable light source switches the light source color in the preset order of white, red, green and blue, and step S6-1 is performed once each time the light source color is switched. A total of 32 images are collected in a single detection; The specific steps of detecting and identifying by the image recognition module using the 32 images collected in a single detection include: Step S6-3, image preprocessing and feature fusion, the image recognition module pre-processes and fuses the 32 multi-light-field image sequences collected in step S6-2, including the following sub-steps: Step S6-3-1, basic image optimization: grayscale conversion is performed on each single-frame image, 5*5 Gaussian filter is used to remove noise, and histogram equalization is used to enhance the contrast between the raw material and the background; Step S6-3-2, three-dimensional topography reconstruction and reflectivity map generation: using the pre-processed multi-angle white light image sequence, through the photometric stereo vision algorithm, according to the bright and dark changes of the same object surface under different directions of light, the surface normal of the object is calculated, and the three-dimensional contour and surface micro-topography of the object are reconstructed. Simultaneously, the algorithm calculates the inherent reflectivity of the object surface, and generates a reflectivity map that eliminates the interference of highlights and projections; Step S6-3-3, multi-spectral feature fusion, including the following sub-steps: Step S6-3-3-1, data alignment and multi-spectral reflectance calculation: pixel-level registration is performed between the multi-spectral image sequence and the white light image sequence, and based on the photometric stereo vision principle, the reflectance images of corresponding wave bands are respectively calculated and generated for the red, green, blue and other spectral images to form a multi-wave band reflectance cube; Step S6-3-3-2, depth feature extraction and fusion: the multi-wave band reflectance cube is input into a pre-trained feature extraction network to generate a high-order feature map that integrates the object's intrinsic color, fine texture and spectral characteristics, serving as a comprehensive image for subsequent identification; Step S6-4, rapid distinguishing of raw materials and foreign matters, including the following sub-steps: Step S6-4-1, multi-dimensional feature vector construction: depth features of color and texture are extracted from the high-order feature map generated in step S6-3-3, and three-dimensional contour data calculated in step S6-3-2 are simultaneously imported to jointly form a multi-dimensional feature vector for comparison; Step S6-4-2, comparison and judgment based on the feature library: the multi-dimensional feature vector constructed in step S6-4-1 is rapidly compared with the food raw material feature library pre-stored in the data storage module; the system judges the target with a comprehensive feature similarity lower than a set threshold as a foreign matter, and judges the target with a similarity higher than the threshold as a food raw material; Step S6-5, fine defect detection and grading of the target judged as a food raw material in step S6-4, including the following sub-steps: Step S6-5-1, defect identification and classification: based on the three-dimensional contour data obtained in step S6-3-2, the morphological defects of the target such as fracture, missing angle and deformation are identified; based on the high-order feature map obtained in step S6-3-3, the surface defects such as mold spot, discoloration and scar are identified; Step S6-5-2, defect grading: mold spots and morphological defects that cause serious structural incompleteness are marked as first-level defects; obvious discoloration, missing angle and surface scars that affect appearance are marked as second-level defects; non-molded slight color difference, tiny spots and slight irregular shape are marked as third-level defects; Step S6-6, final screening based on defect tolerance, including the following sub-steps: Step S6-6-1, calling tolerance rules: according to the raw material type selected in step S1, the defect tolerance level rules pre-set for the food raw material are called from the data storage module; Step S6-6-2, executing intelligent screening: the defect type and level of each food raw material target are compared with the rules to screen out unqualified food raw material targets; Step S6-6-3, integrate all foreign matters determined in step S6-4-2 and all unqualified food raw materials determined in step S6-6-2 into a final rejection target list; Step S6-7-1, the image recognition module obtains the accurate pixel coordinates of each target needing to be rejected from the image and sends them to the data processing module, and the data processing module uses the mapping relationship between the image pixel coordinate system and the space three-dimensional coordinate system pre-calibrated and stored by the system to convert the image pixel coordinates of the target into actual three-dimensional space coordinates on the conveying belt; Step S6-7-2, the data processing module generates specific control instructions of the automatic rejection device according to the actual three-dimensional coordinates and sends them to the linkage control module.

[0012] Further, the flattening mechanism comprises a first screw rod sliding table, a vertical rod, a mounting plate, a first threaded telescopic rod, a pushing plate and a distance detection sensor, the first screw rod sliding table is provided on both sides of the detection section conveying belt, the vertical rod is fixed above the sliding table of the first screw rod sliding table, the mounting plate is installed on the top of the two vertical rods, the pushing plate is installed below the mounting plate through the first threaded telescopic rod, the distance detection sensor is installed at the bottom of the mounting plate, the distance detection sensor is used to detect the distance between the pushing plate and the mounting plate and feed back to the controller, and the data processing module of the controller converts the detected distance into the spacing between the bottom of the pushing plate and the upper surface of the detection section conveying belt.

[0013] Further, the specific process of flattening the raw materials in the raw material compartment under the industrial visual intelligent detection assembly by the flattening mechanism comprises: Step S5-1, the linkage control module receives the flattening instruction; Step S5-2, read the granularity of the current level and obtain the height h corresponding to the level; Step S5-3, the first threaded telescopic rod is started to lower the pushing plate, and the distance detection sensor continuously detects the distance and feeds back to the controller; Step S5-4, the data processing module of the controller converts the detected distance into the spacing between the bottom of the pushing plate and the upper surface of the detection section conveying belt, and controls the first threaded telescopic rod to stop when the spacing is equal to h; Step S5-5, the first screw rod sliding table is started to drive the pushing plate to reciprocate between the two partition plates at least twice.

[0014] Further, the automatic rejection device comprises a second screw rod sliding table, a second screw rod telescopic rod, a vacuum suction assembly and a waste collection bin, the second screw rod sliding table is installed at the bottom of the mounting plate, the bottom of the second screw rod sliding table is provided with the second screw rod telescopic rod, the push rod end of the second screw rod telescopic rod is provided with the vacuum suction tube of the vacuum suction assembly, and the vacuum suction tube vertically points downward; Each layer of the detection section conveyor belt is provided with the automatic rejection device. The process of rejecting impurities and unqualified food raw materials by the automatic rejection device comprises: Step S7-1: The linkage control module adjusts the horizontal coordinates of the vacuum suction tube by controlling the first screw rod sliding table and the second screw rod sliding table according to the coordinates of the impurities and unqualified food raw materials fed back by the data processing module. Step S7-2: The linkage control module controls the push rod extension distance of the first screw rod telescopic rod and the second screw rod telescopic rod, adjusts the height of the vacuum suction tube, and lowers the height of the vacuum suction tube until the vacuum suction tube is attached to the surface of the detection section conveyor belt and holds the unqualified food raw materials or impurities that need to be rejected. Step S7-3: The linkage control module controls the negative pressure generator of the vacuum suction assembly to start, the vacuum suction tube sucks the impurities, and the impurities are transported into the waste collection bin through the pipeline.

[0015] Further, the fixed baffle is provided with a movable baffle driven to open by a motor at the end position of the detection section conveyor belt, the detection section conveyor belt end corresponds to the position of the movable baffle, and a cleaning device is arranged at the position; the cleaning device comprises a cylinder, a push plate, a storage box and a third detection camera; the bottom and both sides of the push plate are provided with brushes; after all batches of single raw materials are transported, the linkage control module opens the movable baffle, controls the detection section conveyor belt to send the raw material compartment to the cleaning position and stops, the push plate is pushed out and retracted under the drive of the cylinder, and the residual raw materials are pushed into the storage box arranged on both sides of the conveyor belt, the brushes can brush the food raw materials adhered to the surface of the conveyor belt, and the third detection camera is used to identify whether there are residual raw materials or dust impurities.

[0016] Further, a collection mechanism is further included, the collection mechanism comprises a plurality of buffer boxes and a collection hopper, the bottom of each buffer box is provided with a switch baffle driven by a motor, and each switch baffle is electrically connected with the controller; the outlet of the collection hopper is located above the confluence section conveyor belt, and the confluence section conveyor belt is connected to a discharge storage tank.

[0017] The beneficial effects of the raw material conveying device for food processing are as follows: through integrated automation design and advanced multi-dimensional visual detection technology, a plurality of technical problems existing in the traditional food raw material sorting process are effectively solved. First, the device realizes full-process automation from feeding, identification, diversion, detection to rejection, completely changing the inefficient, labor-intensive and poor consistency operation mode relying on manual visual screening, greatly improving the production efficiency and sorting accuracy; secondly, aiming at the industry pain point that the optical properties of food raw materials are complex and defects and foreign matters are difficult to accurately identify, multi-light field imaging and photometric stereo vision technology are innovatively adopted, 32 image sequences are collected through switching different color and angle light sources, and three-dimensional topography reconstruction and reflectivity decoupling analysis are combined, which can effectively overcome the high light and shadow interference, significantly improve the detection rate of complex defects such as mold spots, micro cracks, transparent foreign matters and color similar impurities; thirdly, by introducing an intelligent screening mechanism based on defect tolerance rules, the differentiated and refined sorting of different categories of raw materials is realized, the diversified product quality standards are met, and the problems of waste of high-quality raw materials or missed detection of substandard products are avoided; in addition, the thinning mechanism, layered detection layout and collection buffer mechanism designed in the device respectively solve the problems of raw material accumulation affecting visual detection effect, multi-layer conveyor belt detection interference and raw material falling and breaking or splashing, ensuring the stability and reliability of the whole conveying and detection process; finally, the device has good interactivity and maintainability, parameter setting and monitoring can be completed through the touch screen, and the cleaning device can automatically clean the residues, ensuring the continuous cleanliness and long-term stable operation of the equipment. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 It is a structural schematic diagram of the raw material conveying device for food processing of the application; Figure 2 It is a structural schematic diagram of the diversion conveying device of the raw material conveying device for food processing of the application; Figure 3 It is a structural schematic diagram of the detection section conveying belt of the raw material conveying device for food processing of the application; Figure 4 It is a structural schematic diagram of the collection mechanism of the raw material conveying device for food processing of the application; Figure 5 It is a system architecture diagram of the controller of the raw material conveying device for food processing of the application; Figure 6 It is a working flow schematic diagram of the raw material conveying device for food processing of the application.

[0019] Wherein, 1-conveying body, 11-identification section conveyor belt, 12-shunt conveying device, 121-housing frame, 122-slide plate, 123-sieve plate, 13-detection section conveyor belt, 131-baffle, 132-raw material compartment, 133-fixed baffle, 1331-movable baffle, 14-converging section conveyor belt, 2-feeding device, 21-feeding hopper, 22-screw conveyor, 3-discharge storage box, 4-industrial visual intelligent detection assembly, 41-first identification camera, 42-industrial camera, 43-ring RGBW programmable light source, 5-controller, 51-image recognition module, 52-data processing module, 53-linkage control module, 54-interaction module, 55-data storage module, 6-thinning mechanism, 61-first screw rod sliding table, 62-stand rod, 63-mounting plate, 64-first screw telescopic rod, 65-flattening plate, 66-distance detection sensor, 7-automatic rejection device, 71-second screw rod sliding table, 72-second screw telescopic rod, 73-vacuum suction pipe, 8-cleaning device, 81-cylinder, 82-pushing plate, 84-storage box, 85-third detection camera, 9-gathering mechanism, 91-buffer box, 92-gathering hopper. DETAILED DESCRIPTION

[0020] The technical solutions of the present application will be described clearly and completely below in conjunction with the drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0021] As shown in Figures 1-6 The present application is a raw material conveying device for food processing, which comprises a conveying body 1, a feeding device 2, a discharge storage box 3, an industrial visual intelligent detection assembly 4 and a controller 5. The feeding device 2 and the discharge storage box 3 are respectively arranged at both ends of the conveying body 1. The conveying body 1 is composed of a plurality of conveyor belts, including an identification section conveyor belt 11, a shunt conveying device 12, a detection section conveyor belt 13 and a converging section conveyor belt 14. The feeding device 2 comprises a feeding hopper 21 and a screw conveyor 22. The feeding end of the screw conveyor 22 extends into the bottom of the feeding hopper 21, and the discharge end of the screw conveyor 22 is located above the identification section conveyor belt 11. The surface of the detection section conveyor belt 13 is provided with baffles 131 at fixed intervals along the conveying direction. The two sides of each detection section conveyor belt 13 are provided with fixed baffles 133 to block the two sides of the baffles 131, and a raw material compartment 132 is formed between the two baffles 131. The shunt conveying device 12 comprises a shell frame 121, a sliding plate 122 and sieve plates 123, the sliding plate 122 and the sieve plates 123 are arranged obliquely, the sliding plate 122 is fixed at the bottom, a plurality of sieve plates 123 are arranged above the sliding plate 122, the sieve plates 123 shunt food raw materials of different particle sizes, and the sliding plate 122 and the sieve plates 123 convey food raw materials into different detection section conveying belts 13 through slides respectively; The industrial visual intelligent detection assembly 4 is arranged above the detection section conveying belt 13, the thinning mechanism 6 and the automatic rejection device 7 are arranged at the industrial visual intelligent detection assembly 4, and the positions of the industrial visual intelligent detection assemblies 4 of different layers of the detection section conveying belt 13 are staggered; The thinning mechanism 6 is used for thinning food raw materials, so as to avoid that the height of the food raw materials exceeds a threshold value and affects visual detection; The industrial visual intelligent detection assembly 4 comprises a first recognition camera 41, an industrial camera 42 and a mounting bracket, the first recognition camera 41 is arranged above the recognition section conveying belt 11 through the mounting bracket, the first recognition camera 41 is used for recognizing the type and granularity of food raw materials, and the industrial camera 42 is arranged on the detection section conveying belt 13 through the mounting bracket, the industrial camera 42 is used for shooting food raw materials on the detection section conveying belt 13 and recognizing impurities and defective raw materials; The driving motors of the conveying belts, the industrial visual intelligent detection assembly 4, the thinning mechanism 6 and the automatic rejection device 7 are electrically connected with the controller 5, the controller 5 is provided with an image recognition module 51, a data processing module 52, a linkage control module 53, an interaction module 54 and a data storage module 55, the image recognition module 51 is used for detecting and recognizing images shot by the first recognition camera 41 and the industrial camera 42, the data processing module 52 is used for processing and logically judging data, the linkage control module 53 receives signals and drives an execution mechanism, and is used for controlling the operation of the conveying belts and the shooting time of the cameras; the controller 5 is provided with a touch screen for interaction, the interaction module 54 is used for controlling the display of the touch screen and receiving the input of the touch screen, and the data storage module 55 is used for storing food raw material data, a feature library, system pre-calibration parameters and defect tolerance rules, the food raw material data comprises the appearance of qualified food raw materials and the size of qualified food raw materials; The automatic rejection device 7 is used for rejecting impurities and defective raw materials recognized by the industrial camera 42.

[0022] Further, the industrial visual intelligent detection assembly 4 further comprises a ring-shaped RGBW programmable light source 43; the industrial camera 42 transmits the collected image to the image recognition module 51, the ring-shaped RGBW programmable light source 43 can switch four different color light compensation, and can also switch the light source point position to change the local illumination angle, and the linkage control module 53 controls the operation of the ring-shaped RGBW programmable light source 43.

[0023] Further, the working process of the raw material conveying device for food processing includes: Step S1, turn on the system power, select the type of food raw material to be detected on the touch screen, and apply the corresponding food raw material data stored in the data storage module 55; Step S2, pour the food raw material into the feeding hopper 21, and the spiral conveyor 22 conveys the food raw material to above the identification section conveyor belt 11, and the food raw material falls onto the identification section conveyor belt 11; Step S3, the first identification camera 41 takes pictures to identify the type of food raw material, and judges whether it is consistent with the setting, if it is different from the type set in step S1, an alarm is sent through the touch screen, the food raw material is conveyed into the shunt conveying device 12 by the identification section conveyor belt 11, and the screen plate of the shunt conveying device 12 screens the food raw material according to particle size and sends it into the detection section conveyor belt 13 located at different layers; Step S4, when the raw material compartment 132 of the detection section conveyor belt 13 moves below the industrial visual intelligent detection assembly 4, the linkage control module 53 controls the detection section conveyor belt 13 to stop running; Step S5, the thinning mechanism 6 thins the raw material in the raw material compartment 132 that moves below the industrial visual intelligent detection assembly 4; Step S6, the industrial camera 42 of the industrial visual intelligent detection assembly 4 collects images and performs visual detection, the image recognition module 51 detects whether there is impurity or unqualified food raw material, and feeds back its size and coordinates to the linkage control module 53; Step S7, the linkage control module 53 controls the automatic rejection device 7 to reject the impurities and defective raw materials identified by the industrial camera 42; Step S8, the linkage control module 53 controls the detection section conveyor belt 13 to move a fixed distance, so that the next raw material compartment 132 moves below the industrial visual intelligent detection assembly 4, and the raw material compartment 132 that has completed detection moves to the end of the detection section conveyor belt 13, and the food raw material on each layer of the detection section conveyor belt 13 falls into the confluence section conveyor belt 14 and is uniformly conveyed to the discharge storage tank 3.

[0024] Further, the specific process of the industrial camera 42 of the industrial visual intelligent detection assembly 4 collecting images includes: Step S6-1, the lamp beads of the annular RGBW programmable light source 43 are evenly divided into eight blocks at equal angles, and the lamp beads of one block are lit in turn, and one image is collected each time the block is switched; Step S6-2, the annular RGBW programmable light source 43 switches the light source color in the preset order of white, red, green and blue, and step S6-1 is performed once each time the light source color is switched, and a total of 32 images are collected in a single detection; The specific steps of the image recognition module 51 detecting and identifying the 32 images collected in a single detection include: Step S6-3, image preprocessing and feature fusion, the image recognition module 51 pre-processes and fuses the 32 multi-light-field image sequences collected in step S6-2, including the following sub-steps: Step S6-3-1, basic image optimization: grayscale conversion is performed on each single-frame image, 5*5 Gaussian filter is used to remove noise, and histogram equalization is used to enhance the contrast of raw materials and background; Step S6-3-2, three-dimensional topography reconstruction and reflectivity map generation: using the pre-processed multi-angle white light image sequence, through the photometric stereo vision algorithm, according to the light and dark changes of the same object surface under different directions of light, the surface normal of the object is calculated, and the three-dimensional contour and surface micro-topography of the object are reconstructed; simultaneously, the algorithm calculates the inherent reflectivity of the object surface, and generates a reflectivity map that eliminates high light and projection interference; Step S6-3-3, multi-spectral feature fusion, including the following sub-steps: Step S6-3-3-1, data alignment and multi-spectral reflectivity calculation: pixel-level registration is performed on the multi-spectral image sequence and the white light image sequence, and based on the photometric stereo vision principle, the reflectivity map of the corresponding waveband is calculated and generated for the red, green, blue and other waveband spectral images, to form a multi-waveband reflectivity cube; Step S6-3-3-2, depth feature extraction and fusion: input the multi-waveband reflectivity cube into the pre-trained feature extraction network to generate a high-order feature map that integrates the essential color, fine texture and spectral characteristics of the object, as a comprehensive image for subsequent identification; Step S6-4, quickly distinguish raw materials and foreign matters, including the following sub-steps: Step S6-4-1, multi-dimensional feature vector construction: the depth features of color and texture are extracted from the high-order feature map generated in step S6-3-3, and the three-dimensional contour data calculated in step S6-3-2 are simultaneously imported, to jointly form a multi-dimensional feature vector for comparison; Step S6-4-2, comparison and judgment based on feature library: the multi-dimensional feature vector constructed in step S6-4-1 is compared with the food raw material feature library pre-stored in the data storage module 55; the system judges that the target with a comprehensive feature similarity lower than the set threshold is a foreign matter, and judges that the target with a similarity higher than the threshold is a food raw material; Step S6-5, fine defect detection and grading of the target judged as a food raw material in step S6-4, including the following sub-steps: Step S6-5-1, defect recognition and classification: based on the three-dimensional profile data obtained in step S6-3-2, the morphological defects of the target such as fracture, missing angle and deformation are recognized; based on the high-order feature map obtained in step S6-3-3, the surface defects such as mold spot, discoloration and scar are recognized; Step S6-5-2, defect grading: the mold spot confirmed by multi-spectral feature and the morphological defect causing serious structural incompleteness are marked as first-level defects; the obvious discoloration, missing angle and surface scar affecting appearance are marked as second-level defects; the non-molded slight color difference, tiny spots and slight irregular shape are marked as third-level defects; Step S6-6, final screening based on defect tolerance, including the following sub-steps: Step S6-6-1, call tolerance rule: according to the raw material type selected in step S1, the defect tolerance level rule preset for the food raw material is called from the data storage module 55; Step S6-6-2, execute intelligent screening: compare the defect type and level of each food raw material target with the rule, and screen out unqualified food raw material targets; for example: for the rule that the raw material must be complete, any morphological defect of any level is considered unqualified; for the rule that the raw material allows missing angle, only when the missing angle reaches the second or first level defect is it considered unqualified; Step S6-6-3, integrate all foreign matters determined in step S6-4-2 and all unqualified food raw materials determined in step S6-6-2 into a final rejection target list; Step S6-7-1, the image recognition module 51 obtains the accurate pixel coordinates of each target that needs to be rejected from the image and sends them to the data processing module 52, and the data processing module 52 uses the mapping relationship between the image pixel coordinate system and the space three-dimensional coordinate system pre-calibrated and stored by the system to convert the image pixel coordinates of the target into actual three-dimensional space coordinates on the conveying belt; Step S6-7-2, the data processing module 52 generates specific control instructions of the automatic rejection device 7 according to the actual three-dimensional coordinates, and sends them to the linkage control module 53.

[0025] Further, the flattening mechanism 6 comprises a first screw rod sliding table 61, a vertical rod 62, a mounting plate 63, a first screw rod 64, a flat pushing plate 65 and a distance detection sensor 66. The first screw rod sliding table 61 is provided on both sides of the detection section conveyor belt 13. The vertical rod 62 is fixed above the sliding table of the first screw rod sliding table 61. The mounting plate 63 is installed on the top of the two vertical rods 62. The flat pushing plate 65 is installed below the mounting plate 63 through the first screw rod 64. The distance detection sensor 66 is installed at the bottom of the mounting plate 63. The distance detection sensor 66 is used to detect the distance between the flat pushing plate 65 and the mounting plate 63 and feed back to the controller 5. The data processing module 52 of the controller 5 converts the measured distance into the distance between the bottom of the flat pushing plate 65 and the upper surface of the detection section conveyor belt 13.

[0026] Further, the specific process of flattening the raw materials in the raw material compartment 132 below the industrial visual intelligent detection assembly 4 by the flattening mechanism 6 comprises: Step S5-1, the linkage control module 53 receives the flattening instruction. Step S5-2, the particle size at the level is read to obtain the height h corresponding to the level. Step S5-3, the first screw rod 64 is started to lower the flat pushing plate 65. The distance detection sensor 66 continuously detects the distance and feeds back to the controller 5. Step S5-4, the data processing module 52 of the controller 5 converts the measured distance into the distance between the bottom of the flat pushing plate 65 and the upper surface of the detection section conveyor belt 13. When the distance is equal to h, the first screw rod 64 is controlled to stop. Step S5-5, the first screw rod sliding table 61 is started to drive the flat pushing plate 65 to move back and forth between the two partitions 131 at least twice.

[0027] Further, the automatic rejection device 7 comprises a second screw rod sliding table 71, a second screw rod 72, a vacuum suction assembly and a waste collection bin. The second screw rod sliding table 71 is installed at the bottom of the mounting plate 63. The second screw rod sliding table 71 is installed at the bottom of the second screw rod sliding table 71. The second screw rod 72 is installed at the end of the push rod of the second screw rod 72. The vacuum suction tube 73 of the vacuum suction assembly is installed vertically downward. Each layer of the detection section conveyor belt 13 is provided with the automatic rejection device 7. The process of the automatic rejection device 7 rejecting impurities and unqualified food raw materials comprises: Step S7-1: The linkage control module 53 adjusts the horizontal coordinate of the vacuum adsorption pipe 73 by controlling the first screw rod sliding table 61 and the second screw rod sliding table 71 according to the coordinates of the impurities and unqualified food raw materials fed back by the data processing module 52. Step S7-2: The linkage control module 53 adjusts the height of the vacuum adsorption pipe 73 by controlling the extension distance of the first screw rod 64 and the second screw rod 72, so that the vacuum adsorption pipe 73 is lowered until it is attached to the surface of the detection section conveyor belt 13 and holds the unqualified food raw materials or impurities that need to be rejected. Step S7-3: The linkage control module 53 controls the negative pressure generator of the vacuum adsorption assembly to start, and the vacuum adsorption pipe 73 sucks the impurities and transports them to the waste collection bin through the pipeline.

[0028] Further, the fixed baffle 133 is provided with a movable baffle 1331 driven by a motor at the end position of the detection section conveyor belt 13. The movable baffle 1331 is installed on the end of the fixed baffle 133 through a rotating shaft, and the motor drives the rotating shaft to rotate and drive the movable baffle 1331 to rotate. The end of the detection section conveyor belt 13 corresponds to the position of the movable baffle 1331 and is provided with a cleaning device 8. The cleaning device 8 includes a cylinder 81, a push plate 82, a dust suction assembly, a storage box 84, and a third detection camera 85. The bottom and both sides of the push plate 82 are provided with brushes. After all batches of single raw materials are transported, the screw conveyor 22 continues to run, but the first recognition camera 41 cannot detect the feeding, which indicates that the transportation is completed. The linkage control module 53 opens the movable baffle 1331 and controls the detection section conveyor belt 13 to send the raw material compartment 132 to the cleaning position and stop. The push plate 82 is driven by the cylinder 81 to push out and retract, and the residual raw materials are pushed into the storage box 84 provided on both sides of the conveyor belt. The brushes can brush the food raw materials adhered to the surface of the conveyor belt. The side close to the cylinder 81 of the push plate 82 is provided with a plurality of suction ports of the dust suction assembly. The suction ports suck the powdery raw materials or dust remaining on the surface of the conveyor belt. The third detection camera 85 is used to identify whether there are residual raw materials or dust impurities.

[0029] Further, a collection mechanism 9 is further included, the collection mechanism 9 includes a plurality of buffer boxes 91 and a collection hopper 92, the bottom of each buffer box 91 is provided with a motor-driven switch baffle, and each switch baffle is electrically connected with the controller 5; the outlet of the collection hopper 92 is located above the confluence section conveyor belt 14, and the confluence section conveyor belt 14 is connected to the discharge storage tank 3. The food raw materials fall into the corresponding buffer box 91 from the detection section conveyor belt 13, and the switch baffles of the buffer boxes 91 in each layer are opened in turn from top to bottom, so that the food raw materials fall into the buffer boxes 91 in the lower layer in stages, the falling height is prevented from being too high to cause the food raw materials to be broken or rebound, and finally fall into the collection hopper 92, and after collection, the food raw materials are ensured to fall on the confluence section conveyor belt 14 and are concentrated and quickly conveyed to the discharge storage tank 3 arranged at the next food processing point.

[0030] The beneficial effects of the food raw material conveying device are as follows: through integrated automation design and advanced multi-dimensional visual detection technology, a plurality of technical problems existing in the traditional food raw material sorting process are effectively solved. First, the device realizes full-process automation from feeding, identification, diversion, detection to rejection, completely changes the inefficient, high labor intensity and poor consistency operation mode relying on manual naked eye screening, and greatly improves the production efficiency and sorting precision; second, aiming at the industry pain point that the optical properties of the surface of the food raw material are complex and defects and foreign matters are difficult to accurately identify, multi-light field imaging and photometric stereo vision technology are innovatively adopted, 32 image sequences are collected through switching different color and angle light sources, and three-dimensional topography reconstruction and reflectivity decoupling analysis are combined, so that the detection rate of complex defects such as mold spots, micro-cracks, transparent foreign matters and color similar impurities can be effectively improved; third, through the introduction of an intelligent screening mechanism based on defect tolerance rules, the differential and fine sorting of different categories of raw materials is realized, the diversified product quality standards are met, and the problems of waste of high-quality raw materials or missed detection of defective products are avoided; in addition, the thinning mechanism, layered detection layout and collection buffer mechanism designed in the device respectively solve the problems of raw material accumulation affecting visual detection effect, multi-layer conveyor belt detection interference and raw material falling and breaking or splashing, and ensure the stability and reliability of the whole conveying and detection process; finally, the device has good interactivity and maintainability, parameter setting and monitoring can be completed through a touch screen, and a cleaning device can automatically clean residues, so that the continuous cleanliness and long-term stable operation of the equipment are ensured.

[0031] The above describes the present application and its embodiments, which are not limited, and the drawings only show one of the embodiments of the present application, and the actual content is not limited thereto. In general, if a person skilled in the art is inspired by it, without departing from the purpose of the present application, without creative design, similar structural modes and embodiments of the technical solutions are not creative, and should belong to the protection scope of the present application.

Claims

1. A raw material conveying device for food processing, characterized in that, The system includes a conveying body (1), a feeding device (2), a discharge storage box (3), an industrial vision intelligent detection component (4), and a controller (5). The feeding device (2) and the discharge storage box (3) are respectively located at both ends of the conveying body (1). The conveying body (1) consists of multiple conveyor belts, including an identification section conveyor belt (11), a diversion conveyor device (12), a detection section conveyor belt (13), and a confluence section conveyor belt (14). The feeding device (2) includes a feeding hopper (21) and a spiral feeder. The feed end of the screw conveyor (22) extends into the bottom of the feed hopper (21), and the discharge end of the screw conveyor (22) is located above the identification section conveyor belt (11); the surface of the detection section conveyor belt (13) is provided with partitions (131) at fixed intervals along the conveying direction, and fixed baffles (133) are provided on both sides of each detection section conveyor belt (13) to block the two sides of the partitions (131), and a raw material compartment (132) is formed between the two partitions (131). The diversion and conveying device (12) includes an outer frame (121), a slide plate (122), and a sieve plate (123). The slide plate (122) and the sieve plate (123) are inclined. The slide plate (122) is fixed at the bottom. Several sieve plates (123) are provided above the slide plate (122). The sieve plates (123) divert food raw materials of different particle sizes. The slide plate (122) and the sieve plate (123) respectively convey the food raw materials into different detection section conveyor belts (13) through slide rails. The industrial vision intelligent detection component (4) is provided above the conveyor belt (13) of the detection section, and the industrial vision intelligent detection component (4) is provided with a thinning mechanism (6) and an automatic rejection device (7). The spreading mechanism (6) is used to spread food ingredients thinly to prevent the food ingredients from piling up to a height exceeding the threshold and affecting visual detection; The industrial vision intelligent detection component (4) includes a first recognition camera (41), an industrial camera (42) and a mounting bracket. The first recognition camera (41) is mounted on the top of the recognition section conveyor belt (11) via the mounting bracket. The first recognition camera (41) is used to identify the type and particle size composition of food raw materials. The industrial camera (42) is mounted on the detection section conveyor belt (13) via the mounting bracket. The industrial camera (42) is used to photograph the food raw materials on the detection section conveyor belt (13) and identify impurities and defective raw materials. The drive motors of each conveyor belt, the industrial vision intelligent detection component (4), the thinning mechanism (6), and the automatic rejection device (7) are all electrically connected to the controller (5). The controller (5) is equipped with an image recognition module (51), a data processing module (52), a linkage control module (53), an interaction module (54), and a data storage module (55). The image recognition module (51) is used to detect and recognize the images captured by the first recognition camera (41) and the industrial camera (42). The data processing module (52) is used for processing running data and making logical judgments. The linkage control module (53) receives signals and drives the actuators to control the operation of each conveyor belt and the shooting timing of each camera. The controller (5) is equipped with a touch screen for interaction. The interaction module (54) is used to control the display of the touch screen and receive the input of the touch screen. The data storage module (55) is used to store food raw material data, feature library, system pre-calibration parameters, and defect tolerance rules. The food raw material data includes the appearance and size of qualified food raw materials. The automatic rejection device (7) is used to reject impurities and defective raw materials identified by the industrial camera (42).

2. The raw material conveying device for food processing according to claim 1, characterized in that, The industrial vision intelligent detection component (4) also includes a ring-shaped RGBW programmable light source (43); the industrial camera (42) transmits the acquired image to the image recognition module (51); the ring-shaped RGBW programmable light source (43) can switch between four different colors of supplementary light, namely white, red, yellow and blue, and can also switch the light source position to change the local illumination angle; the linkage control module (53) controls the operation of the ring-shaped RGBW programmable light source (43).

3. The raw material conveying device for food processing according to claim 2, characterized in that, The workflow of the raw material conveying device for food processing includes: Step S1: Turn on the system power, select the type of food ingredient to be detected on the touch screen, and apply the corresponding food ingredient data stored in the data storage module (55). Step S2: Pour the food raw materials into the feed hopper (21), and the screw conveyor (22) transports the food raw materials to the top of the identification section conveyor belt (11), where the food raw materials fall onto the identification section conveyor belt (11); In step S3, the first identification camera (41) captures an image, identifies the type of food raw material, and determines whether it is consistent with the setting. If it is different from the type set in step S1, an alarm is issued through the touch screen. The identification section conveyor belt (11) transports the food raw material into the diversion conveyor device (12). The sieve plate of the diversion conveyor device (12) sieves the food raw material according to the particle size and sends it into the detection section conveyor belt (13) located in different layers. In step S4, when the raw material compartment (132) of the detection section conveyor belt (13) moves below the industrial vision intelligent detection component (4), the linkage control module (53) controls the detection section conveyor belt (13) to stop running. Step S5, the spreading mechanism (6) will spread the raw material in the raw material compartment (132) below the industrial vision intelligent detection component (4); In step S6, the industrial camera (42) of the industrial vision intelligent detection component (4) acquires images and performs visual detection. The image recognition module (51) detects whether there are impurities or unqualified food raw materials and feeds back their size and coordinates to the linkage control module (53). In step S7, the linkage control module (53) controls the automatic rejection device (7) to reject the impurities and defective raw materials identified by the industrial camera (42); In step S8, the linkage control module (53) controls the detection section conveyor belt (13) to move a fixed distance, so that the next raw material compartment (132) moves below the industrial vision intelligent detection component (4), and the raw material compartment (132) that has completed the detection moves to the end of the detection section conveyor belt (13). The food raw materials on the detection section conveyor belt (13) of each layer fall into the confluence section conveyor belt (14) and are uniformly transported to the discharge storage box (3).

4. The raw material conveying device for food processing according to claim 3, characterized in that, The specific process for the industrial camera (42) of the industrial vision intelligent inspection component (4) to acquire images includes: Step S6-1: The LED beads of the ring RGBW programmable light source (43) are evenly divided into eight blocks at equal angles. The LED beads of one block are lit up in sequence, and an image is captured once each time the block is switched. Step S6-2, the ring RGBW programmable light source (43) switches the light source color in a preset order of white, red, green and blue. Step S6-1 is executed once for each switch of light source color. A total of 32 images are collected in a single detection. The specific steps of the image recognition module (51) for detection and recognition using the 32 images collected in a single detection include: Step S6-3, Image preprocessing and feature fusion: The image recognition module (51) preprocesses and fuses the 32 multi-field image sequences acquired in step S6-2, including the following sub-steps: Step S6-3-1, Basic Image Optimization: Convert each single-frame image to grayscale, use a 5×5 Gaussian filter to remove noise, and use histogram equalization to enhance the contrast between the raw material and the background. Step S6-3-2, 3D morphology reconstruction and reflectance map generation: Using the preprocessed multi-angle white light image sequence, the photometric stereo vision algorithm calculates the surface normal based on the brightness changes of the same object surface under different directional lighting, and reconstructs the object's 3D contour and surface micromorphology; simultaneously, the algorithm calculates the inherent reflectance characteristics of the object surface and generates a reflectance map that eliminates specular and projection interference. Step S6-3-3, multispectral feature fusion, includes the following sub-steps: Step S6-3-3-1, Data Alignment and Multispectral Reflectance Calculation: Perform pixel-level registration between the multispectral image sequence and the white light image sequence, and based on the principle of photometric stereo vision, calculate and generate reflectance maps for the corresponding bands of the red, green, blue and other spectral images respectively, forming a multi-band reflectance cube. Step S6-3-3-2, Deep Feature Extraction and Fusion: Input the multi-band reflectivity cube into the pre-trained feature extraction network to generate a high-order feature map that integrates the object's essential color, fine texture and spectral characteristics, as a comprehensive image for subsequent recognition; Step S6-4, quickly distinguish raw materials from foreign objects, includes the following sub-steps: Step S6-4-1, Multi-dimensional feature vector construction: Extract the depth features of color and texture from the high-order feature map generated in step S6-3-3, and simultaneously import the three-dimensional contour data calculated in step S6-3-2 to jointly form a multi-dimensional feature vector for comparison. Step S6-4-2, comparison and judgment based on feature library: quickly compare the multi-dimensional feature vector constructed in step S6-4-1 with the food raw material feature library pre-stored in the data storage module (55); the system judges the target with a comprehensive feature similarity lower than the set threshold as a foreign object, and judges the target with a similarity higher than the threshold as a food raw material. Step S6-5 involves performing refined defect detection and grading on the targets identified as food raw materials in step S6-4, including the following sub-steps: Step S6-5-1, Defect Identification and Classification: Based on the three-dimensional contour data obtained in step S6-3-2, identify the morphological defects of the target, such as fractures, missing corners, and deformations; based on the high-order feature map obtained in step S6-3-3, identify surface defects such as mold spots, discoloration, and scars. Step S6-5-2, Defect Classification: Mold spots and morphological defects that cause severe structural incompleteness, as confirmed by multispectral features, are marked as Level 1 defects; obvious discoloration, missing corners, and surface scars that affect appearance are marked as Level 2 defects; minor color differences, tiny spots, and minor irregular shapes that are not moldy are marked as Level 3 defects. Step S6-6, the final screening is performed based on the defect tolerance, including the following sub-steps: Step S6-6-1, call the tolerance rule: according to the raw material type selected in step S1, call the defect tolerance level rule preset for the food raw material from the data storage module (55); Step S6-6-2, Perform intelligent screening: Compare the defect type and level of each food ingredient target with the rules, and screen out unqualified food ingredient targets; Step S6-6-3: Combine all foreign objects identified in step S6-4-2 with all non-compliant food ingredients identified in step S6-6-2 into a final list of removal targets; Step S6-7-1, the image recognition module (51) obtains the precise pixel coordinates of each target to be removed from the image and sends them to the data processing module (52). The data processing module (52) uses the mapping relationship between the image pixel coordinate system and the spatial three-dimensional coordinate system pre-calibrated and stored by the system to convert the image pixel coordinates of the target into the actual three-dimensional spatial coordinates on the conveyor belt. In step S6-7-2, the data processing module (52) generates specific control instructions for the automatic rejection device (7) based on the actual three-dimensional coordinates and sends them to the linkage control module (53).

5. A raw material conveying device for food processing according to claim 4, characterized in that, The thinning mechanism (6) includes a first lead screw slide (61), a vertical rod (62), a mounting plate (63), a first threaded telescopic rod (64), a push plate (65), and a distance detection sensor (66). There are two first lead screw slides (61), which are respectively located on both sides of the detection section conveyor belt (13). The vertical rod (62) is fixed above the slide of the first lead screw slide (61). The mounting plate (63) is installed on the top of the two vertical rods (62). The push plate (65) is installed below the mounting plate (63) through the first threaded telescopic rod (64). The distance detection sensor (66) is installed at the bottom of the mounting plate (63). The distance detection sensor (66) is used to detect the distance between the push plate (65) and the mounting plate (63) and feed it back to the controller (5). The data processing module (52) of the controller (5) converts the measured distance into the distance between the bottom of the push plate (65) and the upper surface of the detection section conveyor belt (13).

6. A raw material conveying device for food processing according to claim 5, characterized in that, The specific process of the spreading mechanism (6) spreading the raw material in the raw material compartment (132) below the industrial vision intelligent detection component (4) includes: Step S5-1, the linkage control module (53) receives the thinning instruction; Step S5-2: Read the granularity of the current layer and obtain the height h corresponding to that layer; Step S5-3: The first threaded telescopic rod (64) is activated, the push plate (65) is lowered, and the distance detection sensor (66) continuously detects the distance and feeds it back to the controller (5). In step S5-4, the data processing module (52) of the controller (5) converts the measured distance into the distance between the bottom of the push plate (65) and the upper surface of the detection section conveyor belt (13). When the distance is equal to h, the first threaded telescopic rod (64) is controlled to stop. In step S5-5, the first lead screw slide (61) is activated, driving the push plate (65) to move back and forth between the two partitions (131) at least twice.

7. A raw material conveying device for food processing according to claim 6, characterized in that, The automatic rejection device (7) includes a second lead screw slide (71), a second threaded telescopic rod (72), a vacuum adsorption assembly and a waste collection bin. The second lead screw slide (71) is installed at the bottom of the mounting plate (63). The second threaded telescopic rod (72) is installed at the bottom of the slide of the second lead screw slide (71). The vacuum adsorption tube (73) of the vacuum adsorption assembly is installed at the end of the push rod of the second threaded telescopic rod (72). The vacuum adsorption tube (73) is vertically downward. Each layer of the inspection section conveyor belt (13) is equipped with the automatic rejection device (7); The automatic rejection device (7) removes impurities and substandard food ingredients through the following process: Step S7-1: The linkage control module (53) adjusts the horizontal coordinate of the vacuum adsorption tube (73) by controlling the first lead screw slide (61) and the second lead screw slide (71) according to the coordinates of the impurities and unqualified food raw materials fed back by the data processing module (52); Step S7-2: The linkage control module (53) controls the extension distance of the push rods of the first threaded telescopic rod (64) and the second threaded telescopic rod (72), adjusts the height of the vacuum adsorption tube (73), and lowers the height of the vacuum adsorption tube (73) until it fits the surface of the detection section conveyor belt (13) and catches the unqualified food raw materials or impurities that need to be removed. Step S7-3: The linkage control module (53) controls the negative pressure generator of the vacuum adsorption component to start, and the vacuum adsorption tube (73) sucks in the impurities and transports them to the waste collection bin through the pipeline.

8. A raw material conveying device for food processing according to claim 7, characterized in that, The fixed baffle (133) is provided with a movable baffle (1331) driven by a motor at the end of the detection section conveyor belt (13). A cleaning device (8) is provided at the end of the detection section conveyor belt (13) corresponding to the movable baffle (1331). The cleaning device (8) includes a cylinder (81), a push plate (82), a storage box (84), and a third detection camera (85). The bottom and sides of the push plate (82) are provided with brushes. After all batches of a single raw material are conveyed, the linkage control module (53) opens the movable baffle (1331) and controls the detection section conveyor belt (13) to send the raw material compartment (132) to the cleaning position and stop. The push plate (82) is pushed out and retracted under the drive of the cylinder (81) to push the residual raw material into the storage box (84) provided on both sides of the conveyor belt. The brush can brush the food raw material adhering to the surface of the conveyor belt. The third detection camera (85) is used to identify whether there is residual raw material or dust impurities.

9. A raw material conveying device for food processing according to claim 8, characterized in that, It also includes a collection mechanism (9), which includes multiple buffer boxes (91) and collection buckets (92). The bottom of the buffer box (91) is provided with a motor-driven switch baffle, and each of the switch baffles is electrically connected to the controller (5). The outlet of the collection bucket (92) is located above the confluence section conveyor belt (14), and the confluence section conveyor belt (14) is connected to the discharge storage box (3).

Images